Call me a skeptic, but I have difficulty envisioning good teaching (or anything good) as the outcome of this program. I see kids with personalities like mine skipping class and going to McDonalds when that teacher concerned about how well we master quadratics goes through it for the 700th time in order to get that pay raise. Some see it as measuring success, I see it as a guarantee for marginalized teaching methods and removing the motivation for good teachers to expand their lesson plan beyond the "standard". As the previous poster mentioned, every good teacher I had helped expand the material to practical, real word applications. Its one thing to spit back an answer to a cross multiplication question, but can you use that skill to find how many more hits you'll need to bat .350 in a game? I try to imagine what it would be like if those same teachers who helped me develop that type of thinking would have instead abandon their effort in order to make sure I could score 2 more questions correctly on the test. Abstract thinking and life skill thinking aren't in lesson plans, but good teachers help students develop those skills. I don't know if that is something you can easily measure, but you realize later on in life that it made the difference between true retention and use-for-the-test-and-forget learning.
But this argument goes to the design of the test. If the test isn't looking at useful skills, then the test should be fixed. I agree that the test should have real-world questions.
This doesn't change the fact that the evaluation of and compensation for a teacher should be centered on how well he teaches, not how long he's been a teacher. There are plenty of teachers who simply do not improve over time.
Only unions continue to preach pay-for-longevity over pay-for-performance. Taxpayers and students both deserve better.